CN102084628B - A traffic manager and a method for a traffic manager - Google Patents

A traffic manager and a method for a traffic manager Download PDF

Info

Publication number
CN102084628B
CN102084628B CN200980124948.5A CN200980124948A CN102084628B CN 102084628 B CN102084628 B CN 102084628B CN 200980124948 A CN200980124948 A CN 200980124948A CN 102084628 B CN102084628 B CN 102084628B
Authority
CN
China
Prior art keywords
queue
meter
reshaper
evaluation
priority
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Expired - Fee Related
Application number
CN200980124948.5A
Other languages
Chinese (zh)
Other versions
CN102084628A (en
Inventor
J·卡尔斯特伦
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Marvell International Ltd
Original Assignee
Marvell International Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Marvell International Ltd filed Critical Marvell International Ltd
Publication of CN102084628A publication Critical patent/CN102084628A/en
Application granted granted Critical
Publication of CN102084628B publication Critical patent/CN102084628B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/50—Queue scheduling
    • H04L47/62—Queue scheduling characterised by scheduling criteria
    • H04L47/6295—Queue scheduling characterised by scheduling criteria using multiple queues, one for each individual QoS, connection, flow or priority
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/10—Flow control; Congestion control
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/10—Flow control; Congestion control
    • H04L47/22—Traffic shaping
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/10—Flow control; Congestion control
    • H04L47/24—Traffic characterised by specific attributes, e.g. priority or QoS
    • H04L47/2458—Modification of priorities while in transit
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/10—Flow control; Congestion control
    • H04L47/26—Flow control; Congestion control using explicit feedback to the source, e.g. choke packets
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/10—Flow control; Congestion control
    • H04L47/29—Flow control; Congestion control using a combination of thresholds
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/50—Queue scheduling
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/50—Queue scheduling
    • H04L47/52—Queue scheduling by attributing bandwidth to queues
    • H04L47/522—Dynamic queue service slot or variable bandwidth allocation
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/50—Queue scheduling
    • H04L47/52—Queue scheduling by attributing bandwidth to queues
    • H04L47/527—Quantum based scheduling, e.g. credit or deficit based scheduling or token bank
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/50—Queue scheduling
    • H04L47/58—Changing or combining different scheduling modes, e.g. multimode scheduling
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/50—Queue scheduling
    • H04L47/60—Queue scheduling implementing hierarchical scheduling
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/50—Queue scheduling
    • H04L47/62—Queue scheduling characterised by scheduling criteria
    • H04L47/6215—Individual queue per QOS, rate or priority
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/50—Queue scheduling
    • H04L47/62—Queue scheduling characterised by scheduling criteria
    • H04L47/622—Queue service order
    • H04L47/6225—Fixed service order, e.g. Round Robin
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/50—Queue scheduling
    • H04L47/62—Queue scheduling characterised by scheduling criteria
    • H04L47/622—Queue service order
    • H04L47/6235—Variable service order
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/50—Queue scheduling
    • H04L47/62—Queue scheduling characterised by scheduling criteria
    • H04L47/625—Queue scheduling characterised by scheduling criteria for service slots or service orders
    • H04L47/6275—Queue scheduling characterised by scheduling criteria for service slots or service orders based on priority
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/70—Admission control; Resource allocation
    • H04L47/80—Actions related to the user profile or the type of traffic
    • H04L47/805—QOS or priority aware
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L49/00—Packet switching elements
    • H04L49/30—Peripheral units, e.g. input or output ports
    • H04L49/3027—Output queuing
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/70—Admission control; Resource allocation
    • H04L47/82—Miscellaneous aspects
    • H04L47/821—Prioritising resource allocation or reservation requests

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

The present invention relates to a traffic manager (1) and a method for a traffic manager (1). The method comprises the step of reading a first data packet (D1-DD) comprised in a first queue (QI-QQ) based on a scheduling priority (SP), the scheduling priority (SP) being determined: - at least partly on a configured priority of the first queue (QI-QQ), - at least partly on a first meter value (MV1-MVM) of a first meter (Mi-MM) associated with the first queue (QI-QQ); and - at least partly on a second meter value (MV1-MVM) of a second meter (M1-MM) associated with a first scheduling node (NI-NN); the first scheduling node being a parent node of the first queue.

Description

Traffic management device and for the method for traffic management device
Technical field
The present invention relates to a kind of scheduler, multiple queue and multiple traffic management device that is graded the scheduling node that is arranged in the one or more dispatch layers under root node of comprising, each scheduling node is configured to serve according to priority queue or the scheduling node of lower dispatch layer.
Background technology
In communication network, the service with many traffic characteristics and quality of service requirement can be shared the bandwidth of physics or logical network interface.The example with the different services that require is voice service, Video service, best effort and control message (control messaging) services.Service can have minimum speed, and this can guaranteed in situation mostly.For example, but in some application (broadband aggregation), the minimum-rate of guaranteeing is by oversubscribe, be that minimum speed can not be guaranteed at every turn, it causes being lowered having very high bandwidth demand service quality in period, for example, has increased the delay of service.
Suppose a kind of tree hierarchy, in this hierarchy, root represents 100Mbps Ethernet interface; Sub-root (children ofthe root) is VLAN (VLANs); And each VLAN has the queue in two these tree middle periods of expression, wherein queue stores belongs to the grouping of service.Further, suppose that each VLAN has the flank speed of 8Mbps, and a described minimum bandwidth with 5Mbps in these two queues ensures.If minimum-rate oversubscribe is not allowed to, 100Mbps Ethernet interface can not be supported the VLAN more than 20, because cross over that the minimum bandwidth of all VLAN ensures and be: 5Mbps*20 <=100Mbps.On the other hand, if oversubscribe is allowed to, can support the VLAN of more than 20.
Because not all user is movable simultaneously, thus oversubscribe can process by statistical multipath conversion, the minimum-rate distributing in scheduler (scheduler) like this and can exceed total admissible rate.
But, therein required minimum-rate and exceed distributed minimum-rate and situation can exist.Even if such situation may only occur with very little probability, but it is desirable to control those situations.
The people's such as Olsen US2005/0249220A1 has described a kind of hierarchical communication amount management system and has guaranteed that each in multiple queues is guaranteed the method that minimum-rate, extra bandwidth are maximally utilised in maximum rate restriction according to predefined weight is shared, each queue can not exceed appointment maximum rate and data link.
In the people's such as Olsen US2005/0249220A1, each queue or node have two groups of attributes; Enter attribute and go out group attribute.Enter how property control packet is entered to queue, and control like this degree of depth of queue.Go out how dequeue of group property control packet, and control like this scheduling of queue with respect to other queues.Further, the people such as Olsen describe the minimum rate propagation that allows child node to be configured with minimum-rate, equate or larger minimum-rate although father node does not have.Pass through minimum rate propagation, father node has minimum-rate with good conditionsi and ensures, mean when having in the child node that minimum-rate ensures while there is the traffic, father node also has and will only ensure for the minimum-rate of the traffic from the child node with this guarantee.
The disclosed minimum rate propagation of the people such as Olsen provides efficiency in application, in this application, oversubscribe is general and can not or do not wish the guarantee to each father node its oneself, but is needed for the transmission of some services that are provided of some child node service.
A shortcoming of the people's such as Olsen the disclosed method and system of US2005/0249220A1 is that priority only propagates into father node and hierarchy, further do not propagate from child node.Therefore, can not with mode accurately process minimum-rate in child node and higher than minimum-rate in father node and situation.Therefore, the people such as Olsen can not process the service of converging and therefore can not control the allocated bandwidth in minimum-rate oversubscribe situation.Another shortcoming is that priority attribute is relevant to the bandwidth of unique user definition, any traffic that reaches this bandwidth is all counted as the priority traffic thus, and be given the priority higher than other lower priority queues, this makes according to used dispatching algorithm, and bandwidth will be distributed between the traffic of high-priority queue.
The people's such as Wu US2007/0104210A1 has described dynamic management buffer memory and minimum and the transmission of maximum shaping data dispatching with the stream to packet in the network equipment, and all output bandwidths can and be utilized fully according to the needs of setting and by justice.
To each queue, ensure during shaping at minimum bandwidth, will, based on minimum bandwidth strict priority register independently, carry out select tape scheduler according to polling dispatching or strict priority scheduling.
After having met minimum bandwidth guarantee, each queue enters maximum bandwidth and allows in region, in this region, scheduler will use weighting difference poll (weighted deficit round robin, WDRR) or strict priority (SP) to select packet from different service quality (QoS) queue.
The method that the people's such as Wu US2007/0104210A1 both openly can not process the service of converging openly can not process yet the service of converging system, and therefore disclosed method and system can not control allocated bandwidth in the situation that of minimum-rate oversubscribe.
Summary of the invention
The object of the invention is to overcome the problem in communication network with minimum-rate guarantee oversubscribe.Especially, the object of the invention is to when needed minimum-rate and be greater than distributed minimum-rate and time, in a controlled manner share bandwidth.In other words, the object of the invention is to the in the situation that of minimum guaranteed rate oversubscribe, one predictable allocated bandwidth means also are flexibly provided.
Another object of the present invention is the minimum-rate that ensures any dispatch layer place.
Described object is reached by traffic management device with for the method for traffic management device, described traffic management device comprises scheduler, several queue and several scheduling node that is hierarchically arranged in one or more dispatch layers, each scheduling node is configured to serve according to priority queue or the scheduling node of lower dispatch layer, and described method comprises the step that reads the first packet that is included in the first queue according to dispatching priority.Described dispatching priority is by the configuration preference level according to the first queue at least in part; Measure evaluation according to first of the first meter relevant to the first queue at least in part; And measure evaluation according to second of the second meter relevant to the first scheduling node at least in part and determine; The first scheduling node is the father node of the first queue.
Embodiments of the present invention are defined in the dependent claims.
Brief description of the drawings
Embodiments of the present invention are described in detail with reference to accompanying drawing, wherein:
Fig. 1 a schematically illustrates according to the structure chart of the traffic management device of an embodiment of the invention;
Fig. 1 b schematically illustrates according to the logical view of the traffic management device of an embodiment of the invention;
Fig. 2 schematically illustrates the example of scheduling grade;
Fig. 3 is schematically illustrated in the scheduling of the queue of A scheduling node;
Fig. 4 is schematically illustrated in without the scheduling of propagating in the situation of priority at the scheduling node of the scheduling node of B, C or P layer;
Fig. 5 is schematically illustrated in the scheduling of propagating in the situation of priority at the scheduling node of B, C or P layer scheduling node;
Fig. 6 schematically illustrates the priority propagation in scheduling hierarchy;
Fig. 7 schematically illustrates the network processing unit of the execution mode that comprises traffic management device of the present invention; And
Fig. 8 schematically illustrates the router/interchanger that comprises one or more traffic management devices.
Embodiment
Referring now to accompanying drawing, the present invention is further described in more detail, identical Reference numeral represents identical or corresponding feature, parts or device in the accompanying drawings.
Fig. 1 a schematically illustrates the structure chart of traffic management device 1 according to the embodiment of the present invention.Fig. 1 b schematically illustrates the logic diagram of traffic management device (TM) 1 according to the embodiment of the present invention.Traffic management device 1 can be arranged in buffer memory, queuing and the scheduling to packet is provided in network system.Traffic management device 1 can be contained in the network processing unit 10 of network system, with reference to Fig. 7.But it also can be arranged to autonomous device, described equipment can be disposed in outside network processing unit and with network processing unit and communicate by letter.
In network system, traffic management device can be arranged to multiple object.For example, traffic management device can be arranged to inflow (ingress) traffic management, outflow (egress) traffic management and virtual output work queue.But, should be appreciated that, in network system, traffic management device can be arranged to other objects.
In inflow and outgoing communication buret reason, traffic management device algorithm should be configured to guarantee to share bandwidth according to service-level agreements.Service-level agreements has for example been described QoS parameter, and this parameter is such as minimum bandwidth, maximum bandwidth, delay, shake and loss probability for receive or send to the network user's the traffic from the network user.
In virtual output work queue, traffic management device is in the input of exchanger arrangement (fabric), according to distributing to the QoS parameter of the packet in queue and in exchanger arrangement or in the bandwidth of output and the availability of data storage of this exchanger arrangement, queuing up and schedule data packets.
Network processing unit can be any of multiple known type; Grouping is drawn together the international patent application no that is incorporated to taking the form of quoting at this as the processor described in PCT/EP2007/055777.This processor can comprise the device of multiple known type; The asynchronous process streamline described in the international patent application no PCT/SE2005/0019696 being incorporated to by reference at this is drawn together in grouping.
Traffic management device 1 is arranged to by one or several input ports 2 and receives packet, and is arranged to according to grouping information and makes the decision of taking what action.Therefore traffic management device 1 should comprise the device that checks 3 for dividing into groups.Packet information can be contained in packet, for example, and in protocol header.Alternately or additionally, grouping information can be used as with the side information of packet and transmit, for example, with by transmitting the entity set attribute of packet to traffic management device.
Further, traffic management device can be arranged to and select packet by the queue being written into.The selection of queue can be according to grouping information.For example, queue number can be contained in grouping information and by reading grouping information, traffic management device is known the queue that packet will be written into.
Traffic management device can also be arranged to according to grouping information, according to queuing message (for example current queue length or average queue length) and/or according to be stored in traffic management device and the parameter relevant to queue, such as for example, drop threshold and parameter for activity queue management (Weighted random early drop (weighted random early discard)), determine it is to abandon packet by drop device 4, or by the device 5 of joining the team, packet is joined the team.
If packet is written in queue Q, it becomes after in queue, any packet has above been gone out team, can be used for dispatching and going out team.Queue can be served by scheduler 6, and scheduler 6 can be served by other schedulers that are in higher level in scheduling hierarchy successively.Scheduler 6 is arranged to the SO service order of determining queue or scheduler node by dispatching algorithm.Going out team's device 7 can be arranged to and make packet go out team.The non-restrictive example grouping of dispatching algorithm comprise first enter first service, time-sharing multiplex, poll, WRR, strict priority queuing, difference poll, difference WRR, Weighted Fair Queuing and the earliest deadline preferential.
Going out team from queue or scheduler node can be by such as leak the traffic shaping devices such as cylinder algorithm or token cylinder algorithm or blocked by pressure (backpressure) signal of uploading receiving by traffic management device temporarily.Uploading pressure signal can receive from the output port of rate-constrained, this output port be almost full for transmitting the local storage of unsettled data.
Going out after team, the output port 8 that packet is passed traffic management device sends.Alternatively, traffic management device carries out reprocessing (post-process) to data before being arranged in and sending.For example, traffic management device can be arranged to editing data packets headers or be arranged as to sideband attribute writing information.
In Fig. 1 b, packet D 1-D denter traffic management device 1 by the data-interface that comprises one or more input ports 2, and with before below described mode is dispatched, be stored in one or more input-buffer Q 1-Q qin, such as one or more queues.After being scheduled, packet D 1-D dfrom queue, read, and exited traffic management device via one or more output ports 8.
Further, traffic management device 1 comprises several one or more dispatch layer L that are hierarchically arranged in 1-L lthe scheduling node N at place 1-N n, each scheduling node N 1-N nbe configured to according to priority service in lower dispatch layer L 1-L lscheduling node N 1-N nor queue Q 1-Q q.
Traffic management device 1 comprises the scheduler 6 of the scheduler of for example classification that can comprise multiple (for example 4) identical scheduler, and each scheduler has single scheduling hierarchy and is assigned to a group communication amount interface; Each scheduler mates its data bandwidth and packet rates.The minimum leaf of scheduling hierarchy is queue Q.For example, scheduling node N in packet multiple by being disposed in (5) layer from queue assembles continuously.Each scheduler can by hierarchical arrangement be neatly taking between the traffic interface of distributing to scheduler, share more than 2048 internal schedule nodes and 8192 queues as feature.Fig. 2 has illustrated to illustrate the example of scheduling hierarchy.
As shown in Figure 2, packet D is stored in 8192 FIFO (first in first out) queue Q, and 4 queues are mapped to the scheduling node N of 2048 A layers ain one.The scheduling node N of 1-1024 A layer abe mapped to the scheduling node N of 512 B layers bin one, and the scheduling node N of 1-512 B layer bbe mapped to the scheduling node N of 128 C layers cin one.The scheduling node N of 1-128 C layer cbe mapped in 16 port P, and 1-16 port P is mapped to scheduler tree T.
As example, can distribute 12 ports to kilomegabit (Gigabit) Ethernet interface at the outflow scheduler of the user side of oversubscribe Metro Ethernet system.In each port, B layer and C layer are corresponding to logic interfacing and service, and A layer is corresponding to user, and queue is corresponding to application.Should be appreciated that, the quantity of layer can be different.Further, if need more shallow degree hierarchy, B layer and C layer can for example be configured to transparent void (dummy) layer, and this void layer is connected to higher or lower level node one to one.
Each scheduling node N a, N band N c, and each port P can comprise strict priority scheduling device, poll (RR) scheduler or difference WRR (DWRR) scheduler.Scheduler tree T can comprise strict priority scheduling device or poll (RR) scheduler.
Further, according to shown in Fig. 1 b and 2, each in queue Q, scheduling node N and port P for example, is associated with dual rate reshaper (two token cylinder reshapers).Rate shaper is bit rate reshaper.
In this manual, dual rate reshaper one of them, be also referred to as the minimum-rate reshaper of minimum (confirmed (commited)) token cylinder reshaper, be called as meter M (for example meter M 1, M 2... M q-1, M qand M p... M m), and another is also referred to as the maximum rate reshaper of maximum (exceeding the quata) token cylinder reshaper, is called as reshaper S (for example reshaper S 1, S 2... S q-1, S qand S p... S mand corresponding reshaper value SV 1-SV qand SV p).
Meter M is configured to define dynamic priority, thus can for example, according to the measurement evaluation MV (MV of meter M 1-MV qand MV p) and meter limit value MLV between relation, between limit priority and lowest priority, dynamically change the priority of relevant queue, node or port.As shown in Fig. 1 b, the first meter M 1there is the meter of being less than limit value MLV 1measurement evaluation MV 1, it provides priority is low.But, if measure evaluation MV 1be equal to or greater than meter limit value MLV 1, priority will be for high.
Reshaper S is configured to the bit rate of the output port 8 of limiting telecommunication amount manager 1.
Be included in the first queue Q 1-Q qin the first packet D 1-D d, from the first queue, read, and be for example allowed to enter the output port 8 of traffic management device 1 according to dispatching priority SP.Dispatching priority SP is by least in part according to the first queue Q being configured 1-Q qpriority CP 1-CP q(with reference to Fig. 3), at least in part according to and the first queue Q 1-Q qthe first relevant meter M 1-M mfirst measure evaluation MV i-MV m, and at least in part according to and the first scheduling node N 1-N nthe second relevant meter M 1-M msecond measure evaluation MV 1-MV mdetermine.
In execution mode, dispatching priority can be further by the first scheduling node N based on being configured at least in part 1-N npriority CP 1-CP ndetermine; The first scheduling node N 1-N nthe first queue Q 1-Q nfather node.
The first and second meter M 1-M mcan be so-called rate shaper and can be provided with any suitable form, for example, with the form of software program or its part or with the form of electronic circuit, optics or the mechanical part of numeral or simulation.Reshaper uses loose or strict token cylinder algorithm, and the license of data is entered based on credit parameter.But entering algorithm, any other suitable license can be used.
Meter M 1-M mbe configured to working as measurement evaluation MV 1-MV mbe equal to or higher than meter limit value MLV (for example MLV 1-MLV m) time high priority layer and when measuring evaluation MV 1-MV mbetween low priority layer during lower than meter limit value MLV, dynamically change its priority layer.
The first and second meter M 1-M mmeter limit value MLV 1and MLV 2be respectively set to 0, the first and second meter and be equal to or greater than 0 and will have high priority at their measurement evaluation, and if their measurement evaluation is less than 0 words, they will have low priority.
If the first packet is read from queue and licensed output port 8, the first and second meter M that enter into traffic management device 1 1-M mfirst and second measure evaluation MV 1-MV mbe reduced corresponding to the first packet D 1-D 2the amount of quantity of position.Alternatively, measure evaluation and can be reduced in addition shaping side-play amount, to future, pre-compensation is carried out in the variation of packet size; For example,, by adding or removing packets headers.Shaping side-play amount can be positive, zero or negative.Shaping side-play amount can single queue or single node configuration or pass to traffic management device as attribute together with grouping.
Further, the first and second meter M 1-M mmeasurement evaluation MV 1-MV mevaluation amount is measured in (each clock cycle of for example processor) increase periodically.Measuring evaluation amount can be set up during configuring, and can be provided with quantity/interval [bits/s].In one embodiment, if the measurement evaluation therefore increasing has exceeded configurable burst sizes B parameter S1-BSM, measure evaluation and be set to BS1-BSM.
Should be appreciated that, in embodiments of the present invention, if the first packet is read from queue and licensed output port 8, the first and second meter M that enter into traffic management device 1 1-M mfirst and second measure evaluation MV 1-MV mcan be increased suitable amount, for example, increase corresponding to the first packet D 1-D dthe amount of quantity of position.
In this embodiment, the priority layer of meter can be height in the time that measurement evaluation is less than or equal to meter limit value, and priority layer can be low in the time measuring evaluation higher than meter limit value.
Further, should be appreciated that, if read from queue and the licensed output port 8 first and second measurement evaluation MV that enter into traffic management device 1 in the first packet 1-MV min the execution mode being increased, the first and second meter M 1-M mmeasurement evaluation MV 1-MV mreduce and measure evaluation amount, each clock cycle of for example processor periodically.
In execution mode, to the first packet D 1-D dread and further to the first packet D to the output port of traffic management device 1-D dlicense enter all right:
-basis and the first queue Q at least in part 1-Q qthe first relevant reshaper S 1-S mthe first reshaper value SV 1-SV m; And
-basis and the first scheduling node N at least in part 1-N nthe second relevant reshaper S 1-S mthe second reshaper value SV 1-SV m.
Can be provided reshaper with any suitable form, for example, with the form of software program or its part, or with the form of numeral or Analogical Electronics, optics or mechanical part.Reshaper uses loose or strict token cylinder algorithm, makes the license of data enter based on credit parameter.But entering algorithm, any other suitable license can be used.
If the first packet D 1-D dby from first queue Q 1-Q qread, with the first queue Q 1-Q qwith the first scheduling node N 1-N nthe first and second relevant reshaper S 1-S mthe first and second shaper value SV 1-SV m, by respectively with corresponding to the first packet D 1-D dposition quantity value reduce.
Further, the first and second reshaper S 1-S mthe first and second reshaper value SV 1-SV mincreased periodically reshaper value amount, each clock cycle of for example processor 2.Reshaper value amount can be set up during configuring, and can be provided with the form of quantity/interval [bits/s].In one embodiment, if the reshaper value therefore increasing has exceeded configurable burst sizes B parameter S1-BSM, shaper value is set to BS1-BSM.
Should be appreciated that, in embodiments of the present invention, if the first packet is read from queue and the licensed output port 8 that enters into traffic management device 1, the first and second reshaper S 1-S mthe first and second shaper value SV 1-SV mcan be increased suitable amount, for example, increase corresponding to the first packet D 1-D dthe amount of quantity of position.
Further, should be appreciated that, if read from queue and licensed output port 8, the first and second shaper value SV that enter into traffic management device 1 in the first packet 1-SV min the execution mode being increased, the first and second reshaper S 1-S mshaper value SV 1-SV mreduce periodically shaper value amount, each clock cycle of for example processor.
In execution mode, to the first packet D 1-D dread and further to the first packet D to the output port of traffic management device 1-D dlicense enter all right:
-priority P the P based on propagated at least in part, this priority P P is corresponding to from lower dispatch layer L 1-L lthe first queue QI-Q qpropagate at higher dispatch layer L 1-L lthe first scheduling node N 1-N ndispatching priority SP.
Further, to the first packet D 1-D dread and further to the first packet D of traffic management device output port 1-D dlicense enter all right:
-priority P the P based on propagated at least in part, this priority P P is corresponding to from being positioned at lower dispatch layer L 1-L lthe first scheduling node N 1-N npropagate into and be positioned at higher dispatch layer L 1-L lfather's scheduling node N 1-N ndispatching priority SP.
The first queue Q 1-Q qthe priority CP of configuration can be strict priority, for example high priority, medium priority or low priority; Or dynamic priority.Fig. 3 has provided and has been configured to have high priority, the queue of medium priority and dynamic priority.
In execution mode, be 3,2,1,0 at the dispatching priority SP of queue layer, DC; Wherein 3 is the highest dispatching priorities, and wherein DC represents " being indifferent to ", and dispatching priority does not affect scheduling decision.
The the first scheduling node N configuring 1-N npriority CP can be strict priority (SPQ) or standard priority.
In execution mode, be 5,4,3,2,1,0 at the dispatching priority SP at node layer place, DC; The 5th, the highest dispatching priority.
But, should be appreciated that, the invention is not restricted to priority, the priority configuring and the dispatching priority that for example provide, and these priority are only used as example consideration.
Fig. 3 schematically illustrates at scheduling node N athe scheduling of the queue at place.According to that illustrates at Fig. 3, queue Q 0-Q 7there is the priority of configuration; Be respectively high, in, in, dynamically, dynamically, dynamically, dynamically, dynamically.Except Q 3each queue comprise packet D.Further, each in queue has and has meter M (for example M that measures evaluation MV 0-M 7) and there is the reshaper S of shaper value SV.Queue Q 0-Q 4and Q 6there is the measurement evaluation (providing with empty meter cylinder in the drawings) that is less than meter limit value, its instruction queue Q 0-Q 4and Q 6there is low priority.Queue Q 5there is with Q7 the measurement evaluation MV that is greater than corresponding meter limit value 5and MV 7(providing with the cylinder of filling in the drawings), its instruction queue Q 5and Q 7there is high priority.
Further, queue Q 0-Q 2and Q 4-Q 6in each there is reshaper S, this reshaper S have be equal to or greater than reshaper limit value reshaper value SV (in the drawings with fill reshaper cylinder provide).Queue Q 3and Q 7there is respectively reshaper S 3and S 7, have respectively the reshaper value SV that is less than corresponding reshaper limit value 3and SV 7(illustrating with empty reshaper cylinder in the drawings).
Meaning as shown, at scheduling node, has the queue Q of the strict configuration preference level of for example high or middle configuration preference level 0-Q 2, as long as queue comprises packet and has shaping token (that is, integer value SV is greater than shaping limit value), dispatched by the configuration preference level according to them.
Further, for the queue with dynamic-configuration priority, meter M and reshaper S are used to dispatch.All dynamic-configuration queues that comprise packet and have the measurement evaluation MV that is greater than meter limit value MLV are scheduled before the queue of dynamic-configuration with the measurement evaluation MV that is less than meter limit value.If the queue of two dynamic-configuration has equal measurement evaluation, the queue with the highest reshaper value can be scheduled before another queue.
As shown in Figure 3, A scheduling node NA can comprise four polls (RR) or Weighted Fair Queuing (WFQ) scheduler and strict priority (SPQ) scheduler.From each the result in four RR and WFQ scheduler be respectively to four strict dispatching priority values: 3,2,1,0 relevant queue Q 0, queue Q 2, queue Q 5with queue Q 4, the 3rd, the highest dispatching priority and 0 be minimum dispatching priority.After strict priority is dispatched, queue Q 0to be selected.
Fig. 4 schematically illustrates dispatching at the scheduling node of B, C or P layer scheduling node, and does not propagate priority.According to illustrated, node N 0and N 2there is configuration preference level strict priority (SPQ) and will therefore be mapped to node N b, N cor the SPQ scheduler of port P and being scheduled with limit priority.
Node N 1and N 3-N 7there is the standard priority of configuration, and will be therefore by respectively according to meter M 1, M 3-M 7measurement evaluation MV 1, MV 3-MV 7dispatch.
Node N 4, N 5and N 7there is the meter of being greater than limit value MLV 4, MLV 5and MLV 7measurement evaluation MV 4, MV 5and MV 7, and will therefore be mapped to node N b, N cor the meter scheduler of port P, meter scheduler is indicated with Min in Fig. 4.
Node N 1and N 6have and be less than respectively meter limit value MLV 1and MLV 6measurement evaluation, and will therefore be dispatched by the reshaper value according to them.Therefore node N 1and N 6be mapped to node N b, N cor the reshaper scheduler of port P, reshaper scheduler is indicated with Max in Fig. 4.Dispatch or WFQ dispatches, node N by RR 2, N 5and N 1strict dispatching priority layer 2,1 and 0 will be mapped to respectively.Thereafter, strict priority scheduling will be carried out scheduling node N according to priority node 2.
Fig. 5 has schematically illustrated to propagate the scheduling of priority to the scheduling node at B, C or P layer scheduling node.As shown in FIG., node N 0-N 7there is the configuration preference level identical with the configuration preference level of the node shown in Fig. 4.Further, the measurement evaluation of the meter of node is corresponding to the measurement evaluation of the meter of the node in Fig. 4.Also be like this for the reshaper value of reshaper.But in Fig. 5, priority is propagated into father node from child node.In the example illustrating, father node N0-N7 has respectively the priority 0,2,1,0,3,3,2,0 of propagation.
Meaning as shown, is configured to have the node N0 of strict priority SPQ and N2 and is independent of and propagates the SPQ scheduler of priority mapping to father node.The node N4 and the N5 that are configured with standard priority and the highest propagation priority 3 are mapped in Fig. 5 in the indicated highest measurement scheduler of Min3, node N 4and N 5have and be greater than respectively meter limit value MLV 4and MLV 5measurement evaluation MV 4and MV 5.
In Fig. 5, do not have node to be configured with standard priority, and there is the measurement evaluation of the propagation priority that is greater than meter limit value and 2 or 1, and therefore do not have node to be mapped to be indicated as the next highest measurement meter scheduler of Min2, and be mapped to the next one-next one (next-next) highest measurement meter scheduler that is indicated as Min1.
Only has a node, node N 7there is standard configuration priority, be greater than meter limit value MVL 7measurement evaluation MV 7, and 0 propagation priority, and it therefore mappedly has a minimum meter scheduler that is designated as Min0 in Fig. 5.
Two node N 1and N 6be configured to there is standard priority, but there is respectively the meter of being less than limit value MVL 1and MVL 6measurement evaluation MV 1and MV 6, and will therefore can not be mapped in meter scheduler Min3, Min2, Min1 or Min0.On the contrary, node N 1and N 6by being mapped to the reshaper scheduler with Max instruction in Fig. 5, because reshaper value SV 1and SV 6be greater than respectively reshaper limit value SVL 1and SVL 6.
By means of RR scheduling or WFQ scheduling, node N 2be selected at N 0on, because N 2there is higher propagation priority, 1 instead of 0.Because node N 2be configured with strict priority, node N 2be given the highest strict dispatching priority value, 5.
By means of RR scheduling or WFQ scheduling, node N 5be selected at N 4on, and shone upon with the highest strict dispatching priority value of the next one (, 4).Further, because node N 7be only by the node with minimum meter scheduler Min0 mapping, therefore it is given next minimum priority value, and 1.
By means of RR scheduling or WFQ scheduling, node N 1be selected at N 6on and be given minimum dispatching priority value, 0.Finally, by means of strict priority scheduling, node N 2be used as highest priority node and dispatch, and its dispatching priority can be transmitted to higher level.
Fig. 6 has schematically illustrated the example that single queue high priority is propagated in scheduling hierarchy.The scheduling of queue, node and port can be according to previous described reaching, and determined dispatching priority can be propagated in hierarchy.The priority of propagating is always to optimize on other node by hierarchy to have higher priority queues and the node of measuring evaluation (for example consistent with minimum (confirmed) token cylinder value), and described other node always only has lower priority query via this hierarchy and has higher priority queues but not consistent with minimum (confirmed) token cylinder value measurement evaluation.
In shown scheduling hierarchy, exist four to propagate priority layer, the principle work of each in these four propagation priority layers based on same.As signal, priority is propagated into A node from queue, propagate into B node from A node, propagates into C node, and propagate into port from C node from B node.But, should be appreciated that, priority propagation can limit according to scheduler, and for example priority propagation can be limited in queue layer and A node layer, but does not allow in B node layer.
Propagate priority and used by node scheduling logic, the node with highest-priority queue can be selected before the node with lower priority queue.
The in the situation that of all having dynamic priority on queue layer and in node layer, as long as measure evaluation higher than limit value (, as long as minimum token exists), propagating priority is visible to scheduling logic.In the time that measurement evaluation is less than limit value,, in the time that minimum token is depleted, node priority is changed to lowest priority.
Strict priority can be implemented for any node in scheduling hierarchy.Strict priority is served with limit priority and is ignored their propagation priority.But they still come (from lower level) based on the mode identical with the node that is configured to have standard priority and propagate their propagation priority.
The dispatching priority of queue is shown in table 1.Dispatching priority is propagated into higher level to propagate priority mechanism.They are by partly for scheduling queue.Label MinTB represents to measure evaluation, and if it is indicated as "Yes", measures evaluation higher than meter limit value, i.e. high priority.If it is indicated as "No", measure evaluation lower than meter limit value, i.e. low priority.Label Max TB represents reshaper value and "Yes"/"No" instruction reshaper value is greater than/is less than reshaper limit value.
Table 1
Data Queue priority Min?TB Max?TB Dispatching priority
* * * High * DC 3
Be In Be DC 2
Be Dynamically Be DC 1
Be Dynamically No * * 0
No DC DC DC DC****
Be High No DC DC
Be In No DC DC
Be Dynamically No No DC
* Min token cylinder is in consistent situation
* Max token cylinder is in consistent situation
* * queue has data transmission
* * * queue is invalid for scheduling
The dispatching priority of node is illustrated in table 2.These priority are by for example, according to configuration preference level and according to token tubular state, calculate according to the reshaper value of the measurement evaluation of correlation-measuring instrument and relevant reshaper.Dispatching priority is by the higher level propagating into propagate priority mechanism.Dispatching priority is by partly for dispatching node.
Table 2
Data Strict priority Min?TB Propagate priority Max?TB Dispatching priority
* * * Be * DC DC 5
Be No Be 3 DC 4
Be No Be 2 DC 3
Be No Be 1 DC 2
Be No Be 0 DC 1
Be No No DC * * 0
No DC DC DC DC DC****
Be Be No DC DC DC
Be No No DC No DC
* Min token cylinder is on consistent state
* Max token cylinder is on consistent state
* * in the tree node always has queue, and wherein data are under it and token
* * * node is invalid for scheduling
The dispatching priority of port is illustrated in table 3.These priority are calculated according to configuration preference level and according to the reshaper value of the measurement evaluation of token tubular state, for example correlation-measuring instrument and relevant reshaper.Dispatching priority is used to port to dispatch.
Table 3
Data Strict priority Min?TB Max?TB Dispatching priority
* * * Be * DC 2
Be No Be * * 1
Be No No Be 0
No DC DC DC DC****
Be Be No DC DC
Be No No No DC
* Min token cylinder is on consistent state
* Max token cylinder is on consistent state
* * in the tree port always has queue, and wherein data are under it and token
* * * port is invalid to dispatching
Fig. 7 has schematically illustrated the network processing unit 10 of the execution mode that comprises traffic management device 1.Anticipate as shown, network processing unit 10 comprises processing unit 18 and is arranged to receive and/or send the data packet interface 12 of packet.For example, interface 12 can be 100Mbps ethernet mac s, kilomegabit (Gigabit) ethernet mac s, 10 Gigabit Ethernet MACs, PCIe interface, SPI-4.2 interface, Interlaken interface etc.
Interface 12 is arranged with traffic management device 1 and communicates by letter with packet buffer 14, for example shared-memory switch.Shared memory switch device 14 is configured to store the packet of one of deviated from network subsystem temporarily, for example interface 12, traffic management device 1 or processing unit 18.Therefore, shared-memory switch 14 can be arranged with interference networks subsystem.
Processing unit 18 can be the data flow streamline with one or more processors of dedicated engine.Processing unit 18 is configured to classification and editing data grouping information, for example head, thus carry out such as functions such as exchange, forwarding and divider walls (firewalling).
Further, for the optional external memory storage 20 of one or more databases, comprise for example forwarding of packets table, can be arranged with network processing unit 10 and communicate by letter.
In the execution mode of network processing unit 10, the interface 12 that packet is passed network processing unit 10 receives.Interface 12 is write grouping the shared-memory switch 14 of buffered packet.Shared-memory switch 14 is write grouping to process the processing unit 18 that divides into groups and the grouping information being used by traffic management device 1 is set.Traffic management device 1 is write in grouping by processing unit 18.Selectable, processing unit 18 is write traffic management device by shared-memory switch 14 by grouping.Further, traffic management device 1 is configured to determine to take what action according to the grouping information of for example being prepared by processing unit 18.If be grouped in traffic management device to enter also going out subsequently team, traffic management device is write interface 12 by shared-memory switch 14 alternatively by grouping, or writes back to processing unit 18 by shared-memory switch 14 alternatively, for reprocessing.Grouping is selectively write back to interface 18 by shared-memory switch 14 from processing unit.Finally, dividing into groups to be passed interface 18 sends from network processing unit 10.
Fig. 8 has schematically illustrated the router or the interchanger 30 that comprise one or more traffic management devices 1, such as one or more inflow traffic management devices 1 and one or more outgoing communication amount manager 2.Router/interchanger 30 can also comprise some network processing units 10.Traffic management device 1 and/or 2 can be contained in network processing unit 10.But, should be appreciated that, network processing unit 10 can be configured to not have traffic management device 1 and/or 2.
As shown in Figure 8, router/interchanger 30 can comprise one or more input ranks of exchanger arrangement 32 to being included in for example bus in router/interchanger 30, and one or more output queue from exchanger arrangement 32.Input rank can be the VOQ 8` of traffic management device.
Should be appreciated that, embodiments of the present invention are configured to ensure minimum-rate at any dispatch layer.For example, scheduling hierarchy can have one group of child node relevant to father node at lower one deck at dispatch layer.Node has the minimum-rate of the measurement evaluation that relates to them and relates to the maximum rate of their reshaper value.The bandwidth of supposing to offer father node be more than or equal to child node minimum-rate and, but be less than child node maximum rate and.In this case, the priority that propagates into child node even all is all 0, is no more than the child node of its minimum-rate, by the priority scheduling of child node with higher than exceeding its minimum-rate, and therefore guarantees the minimum-rate at any layer.
Also ensure in following example by illustration at the minimum-rate of any dispatch layer:
Suppose that the node that is positioned at A layer is equivalent to application corresponding to the network user and queue.Also suppose that each user has at least one application queue of " doing one's best "; For example, for net (web) traffic, but it has maximum rate there is no minimum-rate.Further, suppose that each user has maximum rate and minimum-rate.The available bandwidth that make them and that be no more than the B node layer of distributing to aggregate users by configure user minimum-rate, each user is guaranteed to obtain its minimum-rate.
Although the present invention is described according to the execution mode providing, those skilled in the art will easily find, can change in the case without departing from the scope of the present invention.For example, as described previously, should be appreciated that, in the time that packet is read from queue, measuring evaluation and/or reshaper value can be increased instead of be reduced.Accordingly, be intended to comprise in the above description also should being interpreted as schematically of all the elements illustrated in the accompanying drawings, instead of make an explanation in limiting sense.

Claims (16)

1. the method for the traffic management device of the traffic of supervising the network, described traffic management device comprises hierarchical scheduler, multiple queue and is graded the multiple scheduling nodes that are associated with multiple corresponding dispatch layers, each scheduling node in wherein said multiple scheduling node is configured to serve according to respective priority a scheduling node of the lower dispatch layer in a queue and the described multiple scheduling node in described multiple queue, and described method comprises:
The dispatching priority of the first queue based in described multiple queues, the first packet is read in described the first queue from described multiple queues, the described dispatching priority of described the first queue in wherein said multiple queues based on:
The configuration preference level of described the first queue in described multiple queue,
First of first meter relevant to data in described the first queue in described multiple queues measured evaluation;
Second of second meter relevant to data in the first scheduling node in described multiple scheduling nodes measured evaluation, and described the first scheduling node in described multiple scheduling nodes is the father node of described the first queue in described multiple queue;
The described dispatching priority of described the first queue in described multiple queues is propagated into described the first scheduling node in described multiple scheduling node;
Based on the described dispatching priority that is transmitted to described the first queue in described multiple queues of described the first scheduling node in described multiple scheduling node, change the dispatching priority of described the first scheduling node in described multiple scheduling node; And
The described dispatching priority of described the first scheduling node based in described multiple scheduling nodes, described the first scheduling node from described multiple scheduling nodes reads described the first packet.
2. method according to claim 1, further comprises:
Utilize described the first meter and described the second meter, measuring evaluation and described second and measure the high priority layer of evaluation while being equal to or higher than respectively separately the first meter limit value and the second meter limit value with between described the first measurement evaluation and described the second measurement evaluation low priority layer during respectively lower than described the first meter limit value and described the second meter limit value when described first, dynamically change described the first meter and described the second meter priority layer separately.
3. method according to claim 1, further comprises:
If described the first packet is read by described the first queue from described multiple queues, reduce described first according to the quantity of the position of described the first packet and measure evaluation and described the second measurement evaluation; And
Increase periodically described the first measurement evaluation and described the second measurement evaluation to measure the quantity of evaluation.
4. method according to claim 1, further comprises:
Utilize described the first meter and described the second meter, measuring evaluation and described second and measure between the low priority layer of evaluation during separately respectively higher than the first meter limit value and the second meter limit value and high priority layer in the time that described the first measurement evaluation and described the second measurement evaluation are equal to or less than respectively described the first meter limit value and described the second meter limit value when described first, dynamically change described the first meter and described the second meter priority layer separately.
5. method according to claim 4, further comprises:
If described the first packet is read by described the first queue from described multiple queues, increase described first according to the quantity of the position of described the first packet and measure evaluation and described the second measurement evaluation; And
Reduce periodically described the first measurement evaluation and described the second measurement evaluation to measure the quantity of evaluation.
6. method according to claim 1, further comprises based on the following and reads described the first packet:
The first reshaper value of first reshaper relevant to described the first queue in described multiple queues; And
The second reshaper value of second reshaper relevant to described the first scheduling node in described multiple scheduling nodes.
7. method according to claim 6, further comprises:
If described the first packet is read by described the first queue from described multiple queues, reduce described the first reshaper value and described the second reshaper value according to the quantity of the position of described the first packet; And
Increase periodically described the first reshaper value and described the second reshaper value with the quantity of reshaper value.
8. method according to claim 6, further comprises:
If described the first packet is read by described the first queue from described multiple queues, increase described the first reshaper value and described the second reshaper value according to the quantity of the position of described the first packet; And
Reduce periodically described the first reshaper value and described the second reshaper value.
9. a traffic management device, comprising:
The scheduler of classification;
Multiple queues; And
Be graded the multiple scheduling nodes that are associated with multiple corresponding dispatch layers, each scheduling node in described multiple scheduling node is configured to serve a scheduling node in a queue in described multiple queue and described multiple scheduling node with lower dispatch layer according to respective priority, described traffic management device is configured to the dispatching priority of the first queue based in described multiple queues, the first packet is read in described the first queue from described multiple queues, the described dispatching priority of described the first queue in wherein said multiple queue based on:
The configuration preference level of described the first queue in described multiple queue,
First of first meter relevant to data in described the first queue in described multiple queues measured evaluation;
Second of second meter relevant to data in the first scheduling node in described multiple scheduling nodes measured evaluation, and described the first scheduling node in described multiple scheduling nodes is the father node of described the first queue in described multiple queue;
Described the first queue in wherein said multiple queue is configured to the described dispatching priority of described the first queue in described multiple queues to propagate into described the first scheduling node in described multiple scheduling node; And
Described the first scheduling node in wherein said multiple scheduling node is configured to the described dispatching priority based on being transmitted to described the first queue in described multiple queues of described the first scheduling node in described multiple scheduling node, changes the dispatching priority of described the first scheduling node in described multiple scheduling node.
10. traffic management device according to claim 9, wherein the first meter and the second meter are configured to, measuring evaluation and described second and measure the high priority layer of evaluation while being equal to or higher than respectively separately the first meter limit value and the second meter limit value with between described the first measurement evaluation and described the second measurement evaluation low priority layer during respectively lower than described the first meter limit value and described the second meter limit value when described first, dynamically change described the first meter and described the second meter priority layer separately.
11. traffic management devices according to claim 10, are further configured to:
If described the first packet is read by described the first queue from described multiple queues, reduce described first according to the quantity of the position of described the first packet and measure evaluation and described the second measurement evaluation; And
Increase periodically described first and measure evaluation and described the second measurement evaluation.
12. traffic management devices according to claim 9, wherein the first meter and the second meter are configured to, measuring evaluation and described second and measure between the low priority layer of evaluation during respectively higher than the first meter limit value and the second meter limit value and high priority layer in the time that described the first measurement evaluation and described the second measurement evaluation are equal to or less than respectively described the first meter limit value and described the second meter limit value separately when described first, dynamically change described the first meter and described the second meter priority layer separately.
13. traffic management devices according to claim 12, are further configured to:
If described the first packet is read by described the first queue from described multiple queues, increase described first according to the quantity of the position of described the first packet and measure evaluation and described the second measurement evaluation; And
Reduce periodically described first and measure evaluation and described the second measurement evaluation.
14. traffic management devices according to claim 9, are further configured to read described the first packet based on the following:
The first reshaper value of first reshaper relevant to described the first queue in described multiple queues; And
The second reshaper value of second reshaper relevant to described the first scheduling node in described multiple scheduling nodes.
15. traffic management devices according to claim 14, are further configured to:
If described the first packet is read by described the first queue from described multiple queues, reduce described the first reshaper value and described the second reshaper value according to the quantity of the position of described the first packet; And
Increase periodically described the first reshaper value and described the second reshaper value with the quantity of reshaper value.
16. traffic management devices according to claim 14, are further configured to:
If described the first packet is read by described the first queue from described multiple queues, increase described the first reshaper value and described the second reshaper value according to the quantity of the position of described the first packet; And
Reduce periodically described the first reshaper value and described the second reshaper value with the quantity of reshaper value.
CN200980124948.5A 2008-04-24 2009-04-21 A traffic manager and a method for a traffic manager Expired - Fee Related CN102084628B (en)

Applications Claiming Priority (6)

Application Number Priority Date Filing Date Title
US4768108P 2008-04-24 2008-04-24
US61/047,681 2008-04-24
US61/047681 2008-04-24
SE0800949 2008-04-25
SE0800949-0 2008-04-25
PCT/EP2009/054742 WO2009130218A1 (en) 2008-04-24 2009-04-21 A traffic manager and a method for a traffic manager

Publications (2)

Publication Number Publication Date
CN102084628A CN102084628A (en) 2011-06-01
CN102084628B true CN102084628B (en) 2014-12-03

Family

ID=40810443

Family Applications (1)

Application Number Title Priority Date Filing Date
CN200980124948.5A Expired - Fee Related CN102084628B (en) 2008-04-24 2009-04-21 A traffic manager and a method for a traffic manager

Country Status (4)

Country Link
US (2) US8824287B2 (en)
CN (1) CN102084628B (en)
TW (1) TWI477109B (en)
WO (1) WO2009130218A1 (en)

Families Citing this family (62)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6396819B1 (en) * 1998-03-21 2002-05-28 Richard D. Fleeter Low-cost satellite communication system
US7227884B2 (en) 2000-02-28 2007-06-05 Aeroastro, Inc. Spread-spectrum receiver with progressive fourier transform
WO2007136415A2 (en) * 2005-12-30 2007-11-29 Comtech Mobile Datacom Corporation Mobile satellite communications
WO2009130218A1 (en) 2008-04-24 2009-10-29 Xelerated Ab A traffic manager and a method for a traffic manager
KR101205622B1 (en) * 2008-12-22 2012-11-27 한국전자통신연구원 Apparatus and method for bandwidth allocation of channel
US8548107B1 (en) 2009-01-26 2013-10-01 Comtech Mobile Datacom Corporation Advanced multi-user detector
US9106364B1 (en) 2009-01-26 2015-08-11 Comtech Mobile Datacom Corporation Signal processing of a high capacity waveform
US8615013B2 (en) * 2010-05-18 2013-12-24 Agere Systems Llc Packet scheduling with guaranteed minimum rate in a traffic manager of a network processor
US8675711B1 (en) * 2009-09-25 2014-03-18 Comtech Mobile Datacom Corporation System and methods for dynamic spread spectrum usage
CN101795173B (en) * 2010-01-21 2012-11-21 福建星网锐捷网络有限公司 Method for transmitting downlink data in wireless network and wireless access equipment
JP5604919B2 (en) * 2010-03-18 2014-10-15 富士通株式会社 Packet transfer apparatus and packet transfer method
EP2525534B1 (en) 2011-05-18 2018-04-18 Xelerated NewCo AB Network traffic scheduler and associated method, computer program and computer program product
CN102769583B (en) * 2012-08-03 2015-03-18 深圳中兴网信科技有限公司 User distribution device and user distribution method of instant messaging server
US9330152B2 (en) * 2013-03-18 2016-05-03 Cellco Partnership Grid loader process
US9450881B2 (en) * 2013-07-09 2016-09-20 Intel Corporation Method and system for traffic metering to limit a received packet rate
JP5614482B1 (en) * 2013-09-02 2014-10-29 沖電気工業株式会社 Station side terminal device and path switching method
US9325617B2 (en) * 2013-12-09 2016-04-26 International Business Machines Corporation Overlay capabilities exchange using DCBX
CN104753809B (en) * 2013-12-25 2019-04-02 深圳市中兴微电子技术有限公司 The method and device of token is added in a kind of traffic shaping
EP2894807B1 (en) * 2014-01-10 2019-11-13 Mitsubishi Electric R & D Centre Europe B.V. A method for signalling and determining time and frequency resources to be used in a wireless communications network
KR101881430B1 (en) 2014-05-08 2018-07-24 후아웨이 테크놀러지 컴퍼니 리미티드 Data scheduling method and device
EP2996293B1 (en) * 2014-09-12 2017-08-30 Alcatel Lucent A packet scheduling networking device for deadline aware data flows
US9729439B2 (en) 2014-09-26 2017-08-08 128 Technology, Inc. Network packet flow controller
US10110515B2 (en) * 2014-11-14 2018-10-23 Cavium, Inc. Packet scheduling using hierarchical scheduling process
US10447608B2 (en) * 2014-11-14 2019-10-15 Marvell Semiconductor, Inc. Packet scheduling using hierarchical scheduling process with priority propagation
US10277506B2 (en) 2014-12-08 2019-04-30 128 Technology, Inc. Stateful load balancing in a stateless network
CN105354205B (en) * 2015-01-13 2018-06-29 吴昱珂 Interpersonal relationships management method, interpersonal relationships management system and corresponding intelligent terminal
US9736184B2 (en) 2015-03-17 2017-08-15 128 Technology, Inc. Apparatus and method for using certificate data to route data
EP3076616B1 (en) * 2015-03-31 2017-04-12 Mitsubishi Electric R&D Centre Europe B.V. Method for data traffic management and network node for implementing the same
US9769082B2 (en) 2015-04-01 2017-09-19 Honeywell International Inc. System and method for network bandwidth, buffers and timing management using hybrid scheduling of traffic with different priorities and guarantees
US9769075B2 (en) 2015-04-01 2017-09-19 Honeywell International Inc. Interference cognizant network scheduling
US9762501B2 (en) * 2015-04-01 2017-09-12 Honeywell International Inc. Systematic hybrid network scheduling for multiple traffic classes with host timing and phase constraints
EP3091700B1 (en) 2015-05-05 2018-03-21 Mitsubishi Electric R&D Centre Europe B.V. Method for allocating time-frequency resources for transmitting data packets over a frequency selective channel
US9729682B2 (en) 2015-05-18 2017-08-08 128 Technology, Inc. Network device and method for processing a session using a packet signature
JP2017041808A (en) * 2015-08-20 2017-02-23 富士通株式会社 Communication device, network management device and network system
JP6582731B2 (en) * 2015-08-21 2019-10-02 沖電気工業株式会社 Station-side terminator, subscriber-side terminator, optical communication system, path switching method, path switching program, and wavelength switching method
US9762485B2 (en) 2015-08-24 2017-09-12 128 Technology, Inc. Network packet flow controller with extended session management
EP3136678B1 (en) 2015-08-27 2019-11-27 Tata Consultancy Services Limited System and method for real-time transfer of audio and/or video streams through an ethernet avb network
US9871748B2 (en) 2015-12-09 2018-01-16 128 Technology, Inc. Router with optimized statistical functionality
US9985883B2 (en) 2016-02-26 2018-05-29 128 Technology, Inc. Name-based routing system and method
US10193977B2 (en) * 2016-04-29 2019-01-29 Huawei Technologies Co., Ltd. System, device and process for dynamic tenant structure adjustment in a distributed resource management system
US10205651B2 (en) 2016-05-13 2019-02-12 128 Technology, Inc. Apparatus and method of selecting next hops for a session
US10298616B2 (en) 2016-05-26 2019-05-21 128 Technology, Inc. Apparatus and method of securing network communications
US10257061B2 (en) 2016-05-31 2019-04-09 128 Technology, Inc. Detecting source network address translation in a communication system
US9832072B1 (en) 2016-05-31 2017-11-28 128 Technology, Inc. Self-configuring computer network router
US10200264B2 (en) 2016-05-31 2019-02-05 128 Technology, Inc. Link status monitoring based on packet loss detection
US10841206B2 (en) 2016-05-31 2020-11-17 128 Technology, Inc. Flow modification including shared context
US11075836B2 (en) 2016-05-31 2021-07-27 128 Technology, Inc. Reverse forwarding information base enforcement
US10091099B2 (en) 2016-05-31 2018-10-02 128 Technology, Inc. Session continuity in the presence of network address translation
US10009282B2 (en) 2016-06-06 2018-06-26 128 Technology, Inc. Self-protecting computer network router with queue resource manager
US9985872B2 (en) 2016-10-03 2018-05-29 128 Technology, Inc. Router with bilateral TCP session monitoring
US10425511B2 (en) 2017-01-30 2019-09-24 128 Technology, Inc. Method and apparatus for managing routing disruptions in a computer network
US10778575B2 (en) * 2017-02-20 2020-09-15 Ge Aviation Systems Llc Systems and methods for scheduling a message
EP4195597A1 (en) 2017-03-07 2023-06-14 128 Technology, Inc. Routing device using flow duplication
US10560373B2 (en) 2017-04-06 2020-02-11 Gvbb Holdings S.A.R.L. System and method for timely and uniform distribution for real-time packet transmission
US10432519B2 (en) 2017-05-26 2019-10-01 128 Technology, Inc. Packet redirecting router
US11165863B1 (en) 2017-08-04 2021-11-02 128 Technology, Inc. Network neighborhoods for establishing communication relationships between communication interfaces in an administrative domain
US20190253341A1 (en) 2018-02-15 2019-08-15 128 Technology, Inc. Service Related Routing Method and Apparatus
US11792114B2 (en) 2019-05-23 2023-10-17 Hewlett Packard Enterprise Development Lp System and method for facilitating efficient management of non-idempotent operations in a network interface controller (NIC)
WO2020248166A1 (en) * 2019-06-12 2020-12-17 华为技术有限公司 Traffic scheduling method, device, and storage medium
JP7127743B2 (en) * 2019-07-04 2022-08-30 日本電信電話株式会社 Communication device, communication method and program
US11658902B2 (en) 2020-04-23 2023-05-23 Juniper Networks, Inc. Session monitoring using metrics of session establishment
CN113626161B (en) * 2021-07-09 2023-12-22 中国科学院信息工程研究所 Distributed multi-user data scheduling method and system

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CA2302129A1 (en) * 1999-03-26 2000-09-26 Newbridge Networks Corporation Timestamping arbitration in digital communication devices
US6188698B1 (en) * 1997-12-31 2001-02-13 Cisco Technology, Inc. Multiple-criteria queueing and transmission scheduling system for multimedia networks
CN1581790A (en) * 2003-08-15 2005-02-16 中兴通讯股份有限公司 Multi-stage dispatch method for supporting multi-port multi-business

Family Cites Families (19)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7215637B1 (en) * 2000-04-17 2007-05-08 Juniper Networks, Inc. Systems and methods for processing packets
US20020141427A1 (en) * 2001-03-29 2002-10-03 Mcalpine Gary L. Method and apparatus for a traffic optimizing multi-stage switch fabric network
US7113510B2 (en) * 2002-03-01 2006-09-26 Xinming Allen Lin Hardware self-sorting scheduling queue
US20040125815A1 (en) * 2002-06-24 2004-07-01 Mikio Shimazu Packet transmission apparatus and method thereof, traffic conditioner, priority control mechanism and packet shaper
US7830889B1 (en) * 2003-02-06 2010-11-09 Juniper Networks, Inc. Systems for scheduling the transmission of data in a network device
US7321940B1 (en) * 2003-06-30 2008-01-22 Cisco Technology, Inc. Iterative architecture for hierarchical scheduling
US20050047425A1 (en) * 2003-09-03 2005-03-03 Yonghe Liu Hierarchical scheduling for communications systems
US7477650B2 (en) * 2003-10-02 2009-01-13 Alcatel Lucent Method and apparatus for frame-aware and pipelined hierarchical scheduling
US20050160178A1 (en) * 2003-12-17 2005-07-21 Venables Bradley D. Adjusting shaped rates
US7522609B2 (en) * 2004-01-14 2009-04-21 Cisco Technology, Inc Propagation of minimum guaranteed scheduling rates among scheduling layers in a hierarchical schedule
US7675926B2 (en) 2004-05-05 2010-03-09 Cisco Technology, Inc. Hierarchical QoS behavioral model
US7525962B2 (en) * 2004-12-27 2009-04-28 Intel Corporation Reducing memory access bandwidth consumption in a hierarchical packet scheduler
US7460544B2 (en) * 2004-12-29 2008-12-02 Intel Corporation Flexible mesh structure for hierarchical scheduling
US8730982B2 (en) 2005-11-10 2014-05-20 Broadcom Corporation Scheduling of data transmission with minimum and maximum shaping of flows in a network device
US8130648B2 (en) * 2006-01-04 2012-03-06 Broadcom Corporation Hierarchical queue shaping
US8194690B1 (en) * 2006-05-24 2012-06-05 Tilera Corporation Packet processing in a parallel processing environment
ATE445277T1 (en) 2006-06-22 2009-10-15 Xelerated Ab PROCESSOR AND METHOD FOR A PROCESSOR
WO2009130218A1 (en) 2008-04-24 2009-10-29 Xelerated Ab A traffic manager and a method for a traffic manager
US7986706B2 (en) * 2009-04-29 2011-07-26 Telefonaktiebolaget Lm Ericsson Hierarchical pipelined distributed scheduling traffic manager

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6188698B1 (en) * 1997-12-31 2001-02-13 Cisco Technology, Inc. Multiple-criteria queueing and transmission scheduling system for multimedia networks
CA2302129A1 (en) * 1999-03-26 2000-09-26 Newbridge Networks Corporation Timestamping arbitration in digital communication devices
CN1581790A (en) * 2003-08-15 2005-02-16 中兴通讯股份有限公司 Multi-stage dispatch method for supporting multi-port multi-business

Also Published As

Publication number Publication date
TWI477109B (en) 2015-03-11
US20140369360A1 (en) 2014-12-18
WO2009130218A1 (en) 2009-10-29
US8824287B2 (en) 2014-09-02
US9240953B2 (en) 2016-01-19
US20110038261A1 (en) 2011-02-17
CN102084628A (en) 2011-06-01
TW201001979A (en) 2010-01-01

Similar Documents

Publication Publication Date Title
CN102084628B (en) A traffic manager and a method for a traffic manager
US8917740B2 (en) Channel service manager
EP2695334B1 (en) Packet scheduling method and apparatus
US7158528B2 (en) Scheduler for a packet routing and switching system
CN1989738B (en) Propagation of minimum guaranteed scheduling rates
US7948882B2 (en) Dual leaky bucket flow control method and system
JP6036310B2 (en) Packet switching apparatus, transmission apparatus, and packet scheduling method
US9548872B2 (en) Reducing internal fabric congestion in leaf-spine switch fabric
CN100420241C (en) Information switching realizing system and method and scheduling algorithm
WO2008095397A1 (en) Traffic scheduling method and apparatus thereof
Georges et al. Strict priority versus weighted fair queueing in switched ethernet networks for time critical applications
WO2007038431A2 (en) Scaleable channel scheduler system and method
US7272150B2 (en) System and method for shaping traffic from a plurality of data streams using hierarchical queuing
US7433953B1 (en) Methods and apparatus for reducing arbitration delay in channelized systems by use of additional scheduling hierarchy
CN102082765A (en) User and service based QoS (quality of service) system in Linux environment
CN113014465A (en) Bus transmission interface system based on quality of service characteristics and design method
CN101542985B (en) A packet switching system and a method thereof
Soni et al. Integrating offset in worst case delay analysis of switched ethernet network with deficit round robbin
Lee et al. WF2Q-M: Worst-case fair weighted fair queueing with maximum rate control
Luangsomboon et al. A round-robin packet scheduler for hierarchical max-min fairness
Saha et al. Multi-rate traffic shaping and end-to-end performance guarantees in ATM networks
Kwak et al. A new starting potential fair queuing algorithm with O (1) virtual time computation complexity
Balogh et al. Average delay and queue length model for WRRPQ
Martinez-Morais et al. Providing QoS with the deficit table scheduler
Popescu et al. Dynamic time sharing: a new approach for congestion management

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C41 Transfer of patent application or patent right or utility model
TA01 Transfer of patent application right

Effective date of registration: 20120130

Address after: Stockholm

Applicant after: A strategic position Lelateniuke Co.

Address before: Stockholm

Applicant before: Xelerated AB

ASS Succession or assignment of patent right

Owner name: MAVER INTERNATIONAL LTD.

Free format text: FORMER OWNER: MARVELL ELECTRONIC TECHNOLOGY SWEDEN CO., LTD.

Effective date: 20121219

C41 Transfer of patent application or patent right or utility model
C53 Correction of patent of invention or patent application
CB02 Change of applicant information

Address after: Stockholm

Applicant after: Marvell electronic technology company in Sweden

Address before: Stockholm

Applicant before: A strategic position Lelateniuke Co.

COR Change of bibliographic data

Free format text: CORRECT: APPLICANT; FROM: XELERATED AB TO: MARVELL ELECTRONIC TECHNOLOGY SWEDEN CO., LTD.

TA01 Transfer of patent application right

Effective date of registration: 20121219

Address after: Bermuda in Hamilton

Applicant after: MARVELL INTERNATIONAL Ltd.

Address before: Stockholm

Applicant before: Marvell electronic technology company in Sweden

C14 Grant of patent or utility model
GR01 Patent grant
CF01 Termination of patent right due to non-payment of annual fee
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20141203

Termination date: 20190421